There are multiple paths leading to molecular nanotechnology. Two of the more visible are chemistry and physics.
The chemists at U. Michigan (Choi & Baker) have recently combined dendrimers and DNA to allow directed assembly of more complex structures (here). This is an extension of our previous discussion of DNA based directed assembly methods (here). A memorable quote is by Baker, "So it's like having a shelf full of Tinker Toys."
Now at the same time the physicists and electronics engineers at HP (Kuekes, Stewart & Williams with Heath) are publishing significant advances in molecular electronics with a molecular scale crossbar latch (here, here and here). This technology is based on nanoimprint lithography (and here). They hope to combine this with existing semiconductor methods at the 32nm scale by 2013. The capacity of this technology is in the vicinity of a trillion switches per cm2 which is at least 10,000 times the density of current chips. Methods that likely to plug into existing technologies have a significant advantages by providing incremental improvements in existing industries.
Rumors circulate that behind the scenes that patent(s) may be in preparation for an assembly process that could legitimately be called directed mechanosynthesis (vs. self-assembly, directed-assembly or bulk-assembly (i.e. lithography based methods))1. But one has to ask, "What is the state of parallel mechanosynthesis?" For it is the parallelization of mechanosynthesis that could play a large role in it becoming an important manufacturing process. If that cannot be achieved it would appear that self-assembly or directed assembly (even of large molecules or lacking complete precision) would appear to have advantages. The only other alternative would seem to be that mechanosynthesis has to be extremely fast. Some might say that using mechanosynthesis assemblers can assemble themselves (after all this is what happens in biology). But that fails to take into account the amount of time that nature put into the development of the self-replication process. Lacking a complete self-replicating system the only alternative is a bootstrap process.
Finally, there is biotechnology. It provides all of the benefits of molecular nanotechnology with the possible exception high density of covalent bonds per unit volume. But with respect to parallelization and production costs it is way out in front because it can easily take advantage of self-replication. It has atomic precision and assemblers of many types. The costs of production blueprints (genes) in this arena has recently been significantly reduced by technologies for DNA synthesis using microchips (Gulari, Katz, Church, Gao) (here). The only thing it is lacking is the intelligent design of enzymes. But that similar to the hurdle that the semiconductor industry had to overcome with the semi-intelligent design and layout of chips over the last 20-30 years. It is simpler in some respects (enzymes may contain thousands to tens of thousands of atoms while chips have millions to tens of millions of transistors) but more complex in others (enzymes are 3D structures while semiconductor chips are largely 2D structures.
So asking the question of "Who's on First?" is not unreasonable.
More… Notes (redirected from news item)…
1. From a technical standpoint directed mechanosynthesis is a directed assembly method but in most current cases the directed assembly approaches are based upon non-specific chemical assembly of molecules such as dendrimers or DNA which are combined and then feed into further self-assembly steps. A pure mechanosynthesis method would assemble both the subcomponents and subcomponent aggregates using mechanosynthesis. This tends to reduce the waste due to the lack of precise assembly involved using many chemical assembly methods.